Brief article on benchmarks of Python repository with leading DVCSen

[Nicholas Allen]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


> 
> If revisions are stored in the branches, why would you store them in the
> cache?  

So that if you branch again from a remote location (and you are not in a
shared repository) it won't download the revisions again.

For example (no shared repository used):


bzr branch http://python.org/trunk

then some time later:

bzr branch http://python.org/some-feature

The second command will not download all the revisions again from the
server but use the cache. It will only have to download the unique
revisions for some-feature.

Another example would be if you delete your trunk branch and then you
want the trunk back again. It won't start downloading all of revisions.

Basically the user can forget about it. Shared repositories would still
speed up local branching though but the cache would speed up remote
branching. Users coming from an SVN command will probably think more in
terms of branching a remote location.

Or are you meaning that the branches are remote and the cache is
> local?

No both branch and cache could be local.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkmTMs8ACgkQ1+i51gqqEGl+iACguIKPwLoxtqetA8vpPXuNkb0v
OUoAnR5TatkGEICrpgwxQ8VcOzpWPKlz
=EVXG
-----END PGP SIGNATURE-----